Everyone keeps saying AI will let anyone make a movie. Well, someone actually did it. It took 3,229 individual generations and 242 hours of work.
Not quite the "type a prompt and get a masterpiece" story we've been sold, is it?
So what actually happened?
A filmmaker has used generative AI tools to produce Catacombs — a movie concept that had been shelved for a decade because traditional production would have required a full crew, significant budget, and resources a solo creator simply didn't have. With AI generation tools, one person finally brought the project to life.
But here's the number that should make every designer sit up: 3,229 individual AI generations. That's not "press a button and done." That's thousands of iterations — curating outputs, refining, regenerating, and stitching together a coherent visual narrative from an inherently unpredictable process. The 242 hours of work confirm what many of us suspect but few want to say out loud.
AI-assisted creative work is still work.
The real story isn't "easy." It's "possible."
This is the part of the AI conversation that keeps getting lost in the noise. The genuine shift isn't that AI makes things effortless. It's that AI makes things possible for people who couldn't have done them before.
Ten years ago, if you had a vision for a film requiring elaborate scenes, complex visual compositions, dozens of environments — you needed a team. A budget. Infrastructure. Now, a single person with patience, taste, and a willingness to iterate thousands of times can produce something that would previously have been completely out of reach.
For designers, this matters enormously. We're already seeing the boundaries of "what a designer does" expand into motion, video, brand campaigns, visual storytelling. AI doesn't replace the creative vision — it replaces the production apparatus. You still need the ideas. You still need the eye. You still need to know when something's working and when it isn't.
But you no longer need a production team to execute.
That's a shift that favours designers specifically, because we already have the visual literacy and compositional instincts that separate a good AI output from a mediocre one.
Meanwhile, 75 oil paintings stole the show
Here's a lovely counterpoint. At Annecy this year — the world's biggest animation festival — the breakout hit was Danse Macabre, an animated short made from just 75 oil paintings. Physical canvases, hand-painted, each one serving as a frame in the animation.
No AI. No digital pipeline. Just paint, planning, and an extraordinary amount of craft.
The film captivated audiences precisely because of its tactile, unmistakably human quality. Every brushstroke visible. Every canvas a real object that exists in the world. As AI-generated visuals become more common, the things that can't be faked — physical materials, visible process, the evidence of a human hand — become more magnetic.